An elementary marking scheme, often appropriate for small-scale group work in a controlled environment, has the mark for the group output being allocated equally. Where we wish to encourage and reward task- or process-related activities, individual members can receive different marks. For example, for additive and conjunctive tasks each member's individual work might be identified and graded. For any format of task, individuals may also be graded on various performance criteria. Where there is an overall group mark and these individual grades, we have the choice of adopting an additive or multiplicative model.
With an additive system, the marks for the individuals' performance are added to the group mark. For example, a group report may be marked out of 80 per cent and each individual's contribution out of 20 per cent is then added. Alternatively, the individual's performance may be marked on the basis of penalty points. A group report can be marked out of 100 per cent and the individual's contribution results in deductions starting at 0 per cent for wholly exceptional, -5 per cent for very good, etc. With a multiplicative model, the mark for the group task is adjusted by a weighting given to the individual assessment. Numerous variations are possible, and one way of avoiding the 'grade inflation' endemic with such a system is that the average student in the individual assessment gets the group mark. Those better than average get a percentage above the group mark, and those below average get a percentage below the group mark.
To emphasise the importance of the group work, it is usual to weight it more heavily than the individual mark: for example, 70:30. A group mark can be awarded that the group must then divide amongst themselves. If the scheme works correctly, they will divide the mark between themselves according to their in-depth knowledge of their relative contributions. Back-up procedures in case of non-agreement need to be available.
Consensus over performance assessment is not always achieved and it is important to have agreed arbitration procedures. A University of Nottingham case study in the LTSN Engineering Work Group Report (2002) illustrates the type of flexibility needed. Student groups were asked to reflect on each member's contribution and indicate any mark redistribution they felt appropriate. However, 'if agreement could not be reached then an individual (confidential) peer assessment form was sent to each group member and a peer assessment adjudication constructed by the module staff.'